Strong evidence from family and twin studies demonstrates that major depressive disorder (MDD) is heritable, yet there has been limited progress in identifying the actual genes involved. A separate, perhaps overlapping set of genes is expected to play a role in individual variation in treatment response in MDD. We seek to characterize genetically, using a large set of markers in many genes, patients who differ in their response to standard treatments as defined in the STAR*D protocol. Our initial focus has been two sets of genes most likely to play a role in the etiology of major depression: 1) genes selected on neurobiological grounds because of their known involvement in pathways thought to be important in mood disorders;and 2) genes implicated primarily by their positions within genomic regions implicated by genetic linkage or association studies of major mood disorders and related conditions. These two approaches are complimentary since the neurobiological candidates arise directly from existing etiologic hypotheses while the positional candidates may lead to the discovery of unexpected pathways. Our ultimate goal is to study markers representing every common functional variation in the human genome, once such genome-wide studies become technologically feasible. We utilize state-of-the-art, high-throughput genotyping methods as well as sophisticated methods of genetic analysis that take into account haplotypes and multi-locus interactions in addition to standard, single-marker analyses. Primary comparisons are performed with the group of cases who respond to citalopram and those who do not, but analyses are done in other treatment groups as well, depending on sample size. In the first years of this project, we completed genotyping 738 markers in a set of 68 candidate genes selected by an expert panel. Results implicated several of these genes in treatment outcome and other genes that contribute to adverse effects. For example, we identified a marker in the gene encoding the serotonin 2A receptor (HTR2A) that was reproducibly associated with both response and remission during 6-12 weeks of therapy with citalopram. We also identified markers near two genes encoding the ionotropic glutamate receptors, GRIK2 and GRIA3, that seem to be associated with treatment-emergent suicidal ideation. We also discovered additional markers that predict treatment outcome. One marker is in the gene GRIK4, that encodes yet another ionotropic glutamate receptor. A second marker is in the gene FKBP5, which encodes a protein involved in the trafficking of glucocorticoid receptors, key molecules in the stress response system. This latter finding confirms an earlier report implicating this gene in an independent sample of inpatients with major depression. We also discovered genetic markers that help identify those at risk for sexual dysfunction during antidepressant therapy. Sexual dysfunction, such as erectile dysfunction, is one of the most common complaints during treatment with modern selective serotonin reuptake inhibitor antidepressants. We also expanded our study of treatment-emergent suicidal ideation, and identified additional markers that increase risk of this worrisome event. Taken together with our previous findings, these markers can identify a few individuals at substantially increased risk, who may benefit from closer monitoring or alternative treatments. In the past year, we have developed models that robustly predict antidepressant outcome across independent samples when genetic markers are combined with known clinical predictors. We found that models that incorporate anxiety along with several genetic markers are more predictive than either alone, enhancing our ability to match treatment choice to patients'personal characteristics. Ongoing work aims to expand the search for genetic markers of treatment outcome to the genome-wide level. We have joined a collaboration with investigators from Harvard, the Max Planck Institute Munich, and University College London that aims to carry out a meta-analysis of STAR*D along with the two other genome-wide association studies of antidepressant outcome that have been completed, known as MARS and GENDEP. This combined sample should have greater power to uncover association with common genetic markers. With the support of a K99 award to Gonzalo Laje, we are also commencing a combined MRS imaging and pharmacogenetics study of 100 patients undergoing outpatient treatment for major depressive disorder. With the support of a Bench to Bedside Award, we have completed genotyping of 1000 participants in the HALT-C trial, who recieved alpha interferon for viral hepatitis. We are seeking to find genetic markers of mood and other psychiatric symptoms that commonly emerge during alpha interferon therapy. Finally, we have begun to explore ways in which the power of large-scale sequencing methods can be brought to bear on the problem of highly treatment-resistant depression.